I think most of comparisons between MBTT and socionics, from what I have seen of people making, presenting them, are invalid. Because they either do not recognize or miss the essence of socionics completely. And that's because they tend to come from a jungian perspective. Which is something that is, in reality, optional in socionics, there are soiconics models out there that do not use concepts like Te, Ti at all.
To use an analogy, it's like contrasting a dolphin and a shark superficially. Looking at them they both seem like fish and one can find plenty of superficial similarities on that level. But those are merely the result of them adapting to the same environment. At their core they are immensely different, a dolphin has more in common with an elephant then a shark.
In a nutshell I would say socionics is a study of information production, information exchange and information processing by human beings.
And in order to do that it took and adapted ideas, concepts by Jung. But they are in no way central to it, what makes socionics socionics. I mean, no more then a theory in physics is physics.
And to paint this fundamental difference further, I will attempt to describe what I would say are the central notions of socionics, what makes socionics socionics. From my point of view.
The first premise, where I would say socionics starts from, is that people exchange information.
And then it proceeds to attempt to categorize this information. And it makes what are called informational elements. These are the fundamental categories of information exchanged between people. Much like how atoms are for things.
It then looks at the information exchange between people. What happens when a person receives a certain type of information, that is, how people respond to different information elements. And it attempts to categories these. They are called socionics relations. And they are fundamental categories of information exchanges.
And then it looks at why people respond the way they do. Why some people when presented with certain informational elements respond one way, why others respond another way. And here it introduces the concept of an informational metabolism. Which is basically a claim that people take in information, process it, and emit new information. Where these three steps are linked in a causal link, that is, being presented with a certain informational element will cause a certain reaction. And it attempts to categories these. And they form the fundamental categories of information processing by people (Or otherwise referred to as socionics types).
Right now the most popular socionics model, where a socionics model is a theory that tries to the best of it's abilities to most accurately and completely cover the area of informational production, exchange and processing by people, is the Model A. Which just *happens* to get it's inspiration of how the informational elements should look like and how the informational metabolism should look like from Jung. This does not make socionics itself in any way dependent or associated with Jung. It makes the Model A, but not socionics.